NXP Semiconductors N.V. (NXPI) Earnings Call Transcript & Summary

Good morning, everyone. I'm Vivek Arya, I cover semis and semi cap equipment at BofA Securities. Really delighted you could join us this morning with the session with NXP Semiconductor. Really delighted to have Jeff Palmer, the Senior Vice President of Investor Relations for NXP. And my plan is to go through a number of questions from my side. But if you have anything you want to bring up, please feel free to raise your hand. And towards the end, I'll also pause a few minutes before to see if you have other questions. On that, Jeff...

Thank you. Great seeing you again after a long time.

Likewise. Yes, it's actually our third year conference in person.

So I know a lot of crosscurrents from a macro perspective, Jeff, it will help the audience to kind of just give an overall kind of state of the union, what's going well? I know you guys reported earnings just a few weeks ago. So if you could just help us recap what are the good things you're seeing from NXP perspective, right? What are the macro headwinds that you're facing in your different segments and then we can run through some...

Yes. I think the #1 question we get from kind of a macro perspective and I'm sure in everyone's mind is how is demand and supply going, right? And as you know, in our automotive and industrial business, which is roughly 70% of total revenue, demand is far in excess of supply. And we have been working on the supply side of that equation diligently for 18 months. And the reality is, we don't think it gets anywhere near equilibrium this year '22. We think some of the fundamental challenges of the supply side, which is not a lot of investment in trailing edge technology are going to be with us -- with the industry, not just with NXP, with the industry for some time. We're doing everything we can to drive a successful outcome for ourselves. But it is a challenging environment. I'd say in terms of automotive, it continues to be very strong demand in excess of supply. Gold rates are still below historical norms, right? 2017, we hit 95 million cars being built. Maybe next year, we'll do 80 million, 82 million, I think is what IHS says. So still far below that. In terms of where do we see the most pressure right now in terms of supply, it's actually more in the industrial and IoT business. Automotive, tough, but industrial and IoT is actually much tougher than it has been.

Got it. So Jeff, you have -- you can bring like a very interesting wide perspective of both the company, the industry and the market. There almost seems to be a disconnect where investors and the market think that we are headed into a recession, right, sometime sooner rather than later. But the semiconductor industry is sounding very strong, right, in terms of the demand/supply. What -- why do you think there is this kind of disconnect between the industry and what investors are seeing?

Well, I think part of it is, if you think about how companies -- chip companies think about it, there's more and more content going in everything we focus on: in automotive, in industrial, in mobile, even in data center. The trend towards the secular trends or higher content is clear and pervasive. And we don't see that abating. I think in terms of why do investors think this all going to come to an end? Well, we also in stare at the Bloomberg terminal all day. We see the red numbers all day long. We don't feel good. I think for us, what would worry us from a macro perspective is if you saw consumer demand actually get destroyed. If you had a consumer who said, "Hey, I was going to buy a car in the fall of 2022. But now because rates are far above what they were 2 years ago. I'm going to push that decision out a couple of years." That would be a challenge for us, and we don't see that. I mean, if anything, when we look at all the kind of cookie trails, if you will, we still see very, very low dealer lot inventory, 27, 30 days below the 65-day norm. You still see dealers taking price in terms of benefit above MSRP. You see the auto OEMs shaping their builds towards the higher end of their portfolio, towards premium vehicles and the clear secular trend towards SUVs. So that's driving content for us. Many of those trends are kind of that crafting the build plan towards the premium end, we see in the industrial side of the market as well. So I mean I know you want a joke about this, investors would love us to come up here and sound terrible and say, then we're going to cut numbers or something, but we've actually see business being very strong right now. And we know the concern only too well about this worry that there's excess inventory or something in the channel. Now we don't have a magical system that will tell us exactly how much inventory is at every point in the supply chain. But what we do have is daily interaction with all the top players in that supply chain from the OEMs, the Tier 1s to the Tier 2s. And I got to tell you, we do not hear that there's excess inventory. We're still managing thousands of escalations. Lead times for our products on 80% of our total portfolio are over 52 weeks. And so it's just the facts that we see on the ground don't point to this coming unwound any moment.

Got it. But Jeff, just to kind of push you a little bit on that, which is if, let's say, we remind back 5 or 6 quarters, and I told you, look, in the month of April 22, global SAAR would be like mid-60s. Where do you think NXP's automotive semiconductor business would be? I imagine your number would have been much lower than where it is. So is it just content? Like what else has been the big...

Well, I think the one thing we've talked about this a number of times, you have to go back in history a little bit. If you go back to about the second half of '18 and all through '19, we saw our end markets of auto and industrial being not very good, be kind of weak. It was the first time in 10 years, you had auto sales in China decline. You had our customers who are not just -- they were just not very optimistic. I think a little bit had to do with the U.S.-China trade war. And in hindsight, what we realized what was going on is a lot of our customers were burning off inventory they had, right? As we got to the end of '19 and we all came into the first part of '20, we all walked into the buzzsaw of the pandemic. Any excess inventory there might have been was completely blown out. Simultaneously, our customer struggle canceling orders on us. And as you know, for our business, if you cancel an order, we don't build the product because it's a sole-source type of business. And then subsequently, at the second half of '20, we saw our customers be more optimistic. The future was going to at least continue to be okay. And they placed orders with us and there was really no supply available. And so really, we have been chasing a supply shortage problem since probably the second half of '20. Now there are some secular things that is a content specific xEV, the only 2x the content on an xEV versus an ICE car. If you look at some company-specific drivers that we've talked about since our Analyst Day last year and even before that are coming to fruition and it becoming more material, so that's a positive. I think if you had asked me in 2020, would we be at 75 million, 80 million units this year, I probably would love to say, no, we'll do better than that. It's been a challenge.

Right. Got it. Here you mentioned China, and I know that lockdowns there have impacted, right, consumer demand. Why hasn't that been reflected in the outlook for the auto semiconductor companies?

I can't speak for the auto semi companies. I can tell you for us, when we gave our guidance for Q2, we did derisk the Q2 outlook by several tens of millions of dollars because of potential shutdowns or impacts from shutdowns in China. Now one of the things we want to be very clear on those impacts, we saw more as supply impacts, not demand impacts. We've actually seen a lot of our customers in China, even with shutdowns, continuing to want to maintain the flow of material to them. We're not seeing demand destruction. Now we are seeing some things start to improve incrementally. That's positive. And you guys will see that in the news. And that should have a positive tailwind to the outlook on Q2 guidance.

Got it. Okay. The other headline that we have seen recently is a number of auto OEMs especially, I think both U.S. and European auto OEMs say that they think the supply will be much better in the second half or will improve in the second half. I think they have all said, how does that reconcile with NXP's view that it is still extremely supply constrained.

It's extremely supply constrained versus the demand we see. In terms of, is supply better today than it was last quarter? Is it better than it was last year? Yes. And it will incrementally get better over time. But I think we really want to be clear, there is no big discontinuity moment where it goes from tight to complete equilibrium or even worse complete excess supply. We don't see that occurring.

Got it. Got it. In the past, the semiconductor industry, right, has often discovered problems of excess inventory well passed. It's after the bubble burst. You realize that there was a bubble in the first place. So why will it be different this time? As investors look at inventory rising on balance sheets for many semiconductors. And I'm not asking you to be the spokesperson for the entire industry..

It feels like so.

But now you represent a very important part of it. So people do see inventory rising on balance sheets, right? And at the same time, they see the consumer under pressure in demand. Why isn't that a yellow or red flag?

Is a concern, we're aware of or where have you folks think how investors think and be concerned about excess inventory. And I can tell you our best answer to that is, we stay extremely close to our customers. I mean, we're measuring escalation rates in the thousands. So our general managers are engaging with our customers and our customers' customers on a daily basis. We don't have a magical system that lets us look into their inventory locations and say, how many NXP components do you have? And one of the things that we would think that maybe folks like you on the sell side and certain investors should think about is, we see all the reports on the DOI charts, right? And we've read those reports. You have to remember as you go further along that supply chain, the value of what is being measured or accounted for is higher, right? If you're a car OEM, your DOI is a higher value because you're almost close to the end of building a car. You also have to remember, across the whole supply chain, inflation has been a real and material impact. So your DOI is inflated by that. When we talk to our customers and our customers' customers, we're not hearing that they have excess NXP components. What we're hearing is it is a different part from different vendors, sometimes it's from us. I mean sometimes we're the guys who are short but it's not the same component with every OEM and every point in the supply chain. It's a different problem, week-to-week, day-to-day.

Got it. How do you measure this trend of improving mix. Like, we see auto SAARs, yes, we can get xEV data. But how do you take this next step of saying how many more premium cars, how much more content. Like, is there a way to quantify how much more you're selling to the same customer now versus what you were doing last year?

So we use third-party data for IHS to give us the mix of our xEVs and premium cars. So that's kind of -- we use the same data you use it. Then we look at our own design win momentum. I said we maybe didn't sell 2 years ago or sold 2 years ago in small quantities that we're selling at higher quantity today. We do have a number of new products that are actually quite material now. Radar was a glimmer in our eyes 3 or 4 years ago, $600 million business last year, it will roughly double in the next 3 years. And for electric vehicles, battery management system and other electrical power control type things, very small a few years ago. It's $200 million of business last year, it will double or go at least probably $500 million by '24. The whole move towards a different architecture in the car driving the whole new class of processors that we never sold before. It's actually getting quite material. So the S32 family was roughly $300 million last year. It's going to be $600 million in '24 and actually the real growth of that business is much stronger beyond '24. So these are the things we measure. And you may say, "Jeff, that all are great fine new products, but what about your core business? Do we see things we gain or strengthening there?" We're continuing to see good strong momentum in that part of business too.

Right. Got it. One other thing, Jeff, would love your perspective. You've been -- you add [indiscernible] and now you add NXP so you can bring kind of a very interesting perspective and through different cycles. Is there something different about the semiconductor industry in that? The positive view is that industry is more disciplined, capacity tougher to create a tetra. The negative view is that well, now it is so much ingrained in global GDP so if go through a recession, then semiconductor would really be impacted and lot more. Do you think the industry is really more disciplined sometime?

I do. I mean I've spent my whole career in the semiconductor industry. 30-plus years, I can believe it or not. And it is much more disciplined, it's much more consolidated. There's fewer players. I think the large semi players like ourselves and a number of our peers are actually going to get larger over time. For all the reasons you talk about. Semiconductors are not just nice to have. They are fundamental to many things we do in our lives every day. That doesn't mean it's not a cyclical industry. It is a cyclical industry. I mean we clearly would acknowledge low-end Android handsets are weaker. I think it's pretty obvious. PCs are weaker. But I mean, some lessons are kind of obvious after you went through 2 years of people being stuck at home buying all this since you don't need to buy another PC for your kid, he's got 1, right? So I do think it's more disciplined. I think the industry is much more disciplined in its management of capital and how it deploys capital for building equipment, for building capacity and things like that. I guess I did view that, that way. But it is a cyclical industry. I'm not here to tell you, it's not.

Right. So if, let's say, theoretically, we get into a recession, right, next year, how will NXP behave differently, what will you do differently?

Well, I don't think it's differently. I think we go back to playbooks that I think we've proven we know how to operate very clearly. If we -- let's just take your side of the equation. We do go into recession. What do we do? Things that are discretionary like CapEx, they start to get turned back. Things like incentive compensation gets turned back. We wouldn't be firing engineers because engineers is the lifeblood of the company, and we will continue to maintain the projects they're working on. But anything that's discretionary that we can do to maintain and safeguard the free cash flow for our owners, we'll do. And I think we have a very good track record. NXP is a lot of things. Everybody may have their opinion. But I think the one thing that we've proven is we know how to navigate tougher environments very well.

Got it. Is there a way to quantify pressure on things like gross margins?

Not one that I'm going to say publicly.

Tell me later on.

Yes, I'll tell you later on our earnings call. That's what I'll say. The reality is, Vivek, if you look back over the last 10 years, so today, about 60% of our business comes from the third-party foundry, right? 40% we build internally, that's on the front end. On the back end, packages assembly about 85% is internal. So if you say there is a recession, so all of our foundry business is fundamentally variable costs, right? We don't have the same fixed cost we had a decade ago. I think our fixed cost fixed to variable in aggregate total company. Fixed is about 30%, 35%. And interesting variables, 65%, 70% or something in that area. So you do your math. You say, okay, I'm going to haircut it 10% across the board. And you're smart, I actually in your models. You can kind of figure it. I think the secondary question maybe more interesting is, what end market causes the weakness, right? And is that end market from NXP perspective is, let's say, mobile, just picking mobile as an example, right? That's fundamentally a variable cost business. We build that product of foundries. Now we're not going to do anything that's going to damage our relationship with our foundry partners. But you, as an analyst doing the model can kind of get a sense of, all right, 10% here, what does it do?

Got it. The follow-up to that, Jeff, would be that if you talk with the foundries, right, GlobalFoundries or TSM or whoever, they will say, well, we have years of unlimited visibility, backlogs, long-term agreement, this or the other. But you are saying that relationship is very flexible. And in their minds, they put rate as a lot more fixed in terms of units and pricing.

I think, Vivek, I'm not trying to tell you it's not fixed or not variable. Everything is negotiable, right? Everything is. Like you've heard us start on earnings calls, we have a very, very strong backlog of noncancelable, nonreturnable orders from our customers. And we've matched those in CNRs with commitments we're making to the companies you just mentioned so that we don't have an asymmetric risk problem. But as you know, you've be pragmatic, everything is negotiable in life. And the last thing you want to do is do something that's going to damage your long-term relationship with your customers.

NCNR, do you think have flexibility?

In the aggregate, I think everything is negotiable.

Got it. So it's flexible contract, flexible return of CFR?

A new term you deal. I guess maybe I would say this -- and I know it's a tough job you have veto want to talk in absolute and black and way. Business doesn't -- especially in our business, our industry, and it doesn't run on a spreadsheet right? It's not nice in the, little column, X dollar here is decremental there. It's a much more complex business.

Got it. Okay. The next thing I wanted to ask about, Jeff, is just going into the different end markets. This trend towards increasing content in automotive. Do you think it's possible the auto industry is going to be permanently in this level of much lower levels of SAAR, but much higher levels of -- is it possible we never get back to those 90-plus million units that is -- because the auto industry seems to be comfortable in this range, right?

Well, I get back to absolute. I don't think it's -- is it possible to get back to $95 million? Sure. But you need enough supply to get there. It's going to take a while to get back there. I agree with you. I think the OEMs do like how they've crafted there the mix of their portfolios towards the premium. Fundamentally, xEV vehicles have higher content than their equivalent ICE vehicles. And I think the car OEMs are more -- they like that business. I think the business is changing and that we get us sometimes do you think we're ever going to go back to an environment where you can go on to a dealer a lot and buy a barebones car? I don't think so. I think what used to be a mid-tier car is now a much higher class vehicle. When you go to buy a car, there are fewer options. The cars come out with a higher kit level just to begin with. And yes, they're more expensive. And I think some of the things that are driving that content, safety, right, Radar, cameras, those type of things. Customers value safety. I think the trends that are driving the adoption of xEVs are actually beyond just consumer demand. I think it's governmental demand for things like making the environment safer, better, right? So those things are tailwinds that aren't going to abate next year or the year after. We are here for a long period of time.

Got it. One other thing, Jeff, on staying on this subject of autos is that we saw for a few years, right, in kind of the '16, '17, '18 type of time frame when NXP's growth was sort of lagging the peers, right, for all kinds of internal and external reasons. But over the last 2, 3 years, it's very interesting that even though EVs have taken off, that NXP's growth has kind of been right in line, right, where the best. How has that happened? Because the perception that you guys are a lot less, you're not supplying the power semis where there is that big growth, right? So how has NXP managed to...

Well, first off, I think -- and then I'm going to say our book here for a second, let me. We've made some significant investments in R&D and new programs and our design to revenue cycles, right? From the time that we greenlight a program in R&D until we actually see a product can be 2 or 3 years. From the time we win a new design until it actually goes to revenue, it could be 2 to 3 years. And then the life cycles are 5 to 7 years. So if you think about the problems that we look at are literally decade windows of time. And I think a little what you're seeing right now and is going to seem like ancient history, you're actually finally starting to see the benefits of the merger of NXP and Freescale. I mean really, if you think about this to that, it seems like a long time ago, but it was early '16, right? And if you think about new projects that got green lit at that time, '16-'17, they're just now coming out and into production. So I think we have a better product set. That's one. I think that the company just I think, is in a better place from an opportunity space. I think the market is kind of coming to us and we've invested in the right things.

Right. But do you think it is a fair pushback that, let's say, EV's stake over the world? That companies who are supplying more power semis would be able to grow faster? Or is it a very narrow view of looking at the content...

I think it's a narrow view, but it's a fair view. I think I fully agree that if you look at the bomb of the ICE vehicle versus equivalent xEV, the largest delta in the bomb is power discretes. So yes, those folks who are in that part of the business, good for them. And you might say, why is NXP in the power discrete marketplace? Well, first off, very CapEx-intensive business. We think it's a lower gross margin business. We think there's enough players chasing it that there could be some pricing pressure. It's just not our cup of tea. Where will we focus on that move on the powertrain is on the power management side, battery management, inverters, DC-to-DC converters, right? And we think we have a unique story to tell there. But then let's set the power train aside for a second and look at the rest of the car, the shell. And you look at xEVs versus ICE, actually, xEV has more content just of the fundamental things that you know and love about NXP. But because the OEMs are crafting those products to be more attractive to consumers.

Right. And power is at a premium. So everything else, right, needs to be higher spec than otherwise. No, I think you bring up an interesting point that we can talk about products, but it's also true that NXP is actually the most CapEx efficient among auto suppliers that we follow, right? So I think that's very interesting also. Among these different areas, right, Radar, ultra-wideband, battery management systems, which is the largest and the most exciting in terms of the growth for the next 3, 4 years for NXP?

Well, I think the largest today is Radar, right, from an automotive perspective $600 million last year, right? I mean that's quite material.

How does the unit content type -- I love unit content...

In the best of that. I think in Radar, it's a very simple story. It's kind of a 3-way multiplicator story. Every year, there's more cars with Radar, there are more Radar nodes per car. And as the OEMs ask for more features per node, we have to spend more money and get higher content. So it's that multiplicative effect. That's what's driving the tailwind. Now it's not like this is a dream. This is a $600 million business, and we have the next number of years already underwritten with design awards. So for us, it's really execution. So that's the largest. From an exciting perspective, and this is the engineer in me coming back out as we talked about at our Analyst Day this whole move from flat automotive architectures to more hierarchical domain and zonal architectures, and our play there is a new product family called the S32 processor family, both for domain processors. These aren't microcontrollers. These are pretty big multi-core processors and zonal processors. And I think that's really exciting. And that business last year was $300 million, and that's -- think of that as very, very small in terms of the long-term trajectory. That will double to about $600 million by '24. But the exciting part is where this architectural shift will go close '24.

Got it. Do you worry about the competitive situation, right? We have seen Qualcomm by Veoneer as an example, right? NVIDIA is going really hard after the autonomous side. Ambarella, I think they announced their domain. So is it becoming a more crowded industry?

I don't think it's crowded. I think there's enough space for a number of players. The folks who you mentioned all focus on different parts. I think Qualcomm and NVIDIA are great companies are focused more on kind of that safety, ADAS, fusion processor part of the marketplace of the domains. We've identified with our partners 5 different domains. We think we can be very successful in probably 3 of that hands down. We can be very competitive in the fourth. And I think we'll do okay in that fusion space Because a lot of these guys are coming from the mobile handset market or from the graphics card market into that space are commercial semi-vendors. And the OEMs still want safety and quality. So there's a lot of these designs that have coprocessors to provide safety co-processing. We can do very well in that space.

Got it. I have a few other product questions, but given the time I wanted to ask a few on the financial side. So on the gross margin side, right, NXP is approaching kind of the high end, right, of the range. But at the same time, we see other companies in the industry, a Microchip is like 10 points higher gross margin. So how is it that 2 players relatively similar, not the same, such a large difference in margins?

Well, I think, look, first off, I know it's very easy to just think that we make 1 part and you'd say that part has this gross margin. The number of parts we build and design are measured in the thousands. And so internally, I think we're 48 different business units. Each business unit has its own portfolio. Those portfolios range some older products and midlife products and some newer products. There's different gross margins. So there's a lot of interest segment mix. That's one. Secondly, we do have a view, you can disagree, but we believe growth -- top line growth is worth more than just grinding out a few more 100 basis points of gross margin. Now that's not to say that we don't think it's important. Probably beginning in 2017, we started to focus on making sure that in aggregate on average, the gross margin, new product gross margins are always accretive, right? So we don't greenlight highly dilutive programs.

Right. Got it. You ever see a scenario where NXP's gross margin start with [ 6% ]?

I would love to see that. But I think we have no fundamental reason to dislike that. It's just -- but we're not going to sacrifice the growth in the business and sacrifice the investments we make just to satisfy that number.

I see. So that's the trade-off.

We trade off. I mean there's been a number of financial analysis models to show you for every 100 basis points of more growth, it has a better value to your equity than a few 100 basis points in gross margin. I know that's heretical.

Right. The next thing is on pricing, right? The industry has benefited a lot. And part of that has been predicated on, well, my foundry is raising their price to me, so I have to pass it along. Do you see a situation where foundries don't raise prices and then the semi industry has to roll back their pricing also?

Well, I think you portrayed it well. The reason we are raising prices at least at NXP is inflationary input costs. And all we are doing is passing along the increased inflationary costs to maintain our financial model. We're not padding the prices to take price to make our margins look better. We could, but we decided that's not what we want to do. And what we've heard from our customers, this will pay dividends to us long term. In terms of design awards, becoming larger players to our customers. So we think it's valuable. So to your question about what happens on the other side. Could the foundries start to say, look, we're going to stop raising prices? Okay. But I don't think we retrace all the way back to where pricing or costs were 18 months ago. We may get to a point at some point in the future where we'll start to extend annual price reductions like we have in the past. I think there'll probably be a much smaller amount and they'll be from that higher level. You're not going to retrace all the way back to where your prices were 18 months ago.

Got it. So the historical price negotiation, right, so it will not be at the same level as the kind of discounts in the industry has had to give in the past?

Yes, I'll give you example. So when I started at NXP, I think our annual price takedowns in automotive on average were low single digits, 3% to 5%, let's just say that. I'd say before the pandemic, our price takedowns are probably 100 to 200 basis points/and so that's kind of the new range you'd probably think about off of a higher baseline. And you have to also think about how you win awards in automotive and industrial. You agree you get an award and you agree to a price curve over the lifetime of the program.

Right. Got it. Makes sense. Let me pause there and see if there are any questions on the audience. No. Okay. The next thing is on the industrial IoT business. How much of that is how we should think about as kind of pure industrial versus IoT starts to get a little bit more into a little bit of the consumer side also. So how much of that -- and the growth rates you expect from that business are, I think, almost as strong as what you expect...

The same growth rates, 9% to 14%. So about 60% of our industrial and IoT business is traditional industrial, factory automation, things like that, building automation. 40% is at IoT. And I agree with you. IoT is an amorphous kind of a definition, right? It can range from some wearables to a lot of smart home stuff that we do.

Right. Got it. So isn't that going to be under pressure as the consumer scales back?

It depends with that. I mean I'm not going to give you a black and white answer. We've not seen it as of yet.

Okay. Is it that WiFi is in -- is there a shortage of WiFi? Is that the reason why?

Well, there is a shortage of is foundry capacity, the largest part of that business is processors. Microcontrollers, application processors, crossovers, that's the largest piece. Second largest piece is analog mixed signal, which we sell is kind of our long attached. Third after that is things like connectivity and security. That's our story in industrial selling all those products together. It's not like, oh, the shortage is just WiFi. The shortage is third-party foundry capacity for those products.

Got it. So if we dig down into third-party, so I know you don't -- I think a very small part of the portfolio uses like the most leading edge, right? I do know that NXP has like some 7 or 5-nanometer?

We have a 5-nanometer project for automotive that will sample in '24. And we have the S32 family is on 16-nanometer FinFET today. But -- and this is the majority of our business that we buy externally is 28, 40, 55, some 60 and 90 nanometer. And this sounds crazy, but we have 55-nanometer products that are just starting to ramp into volume and will ramp for quite some time. And that part of the market that's truly trailing edge is not something a lot of the large foundries are overly interested investing in.

Got it. Why is that? Because one would think that, that is not a very high capital burden. Why aren't they interested? If they see the demand.

I think certain foundries -- do you see an opportunity and they are investing like global. I think some foundries like TSM, who's really the King Kong in the marketplace, they are much more focused on data center, mobile and other things. And automotive and industrial is actually a small part of their business. But I think you'd have to ask them as to why they're not investing more there.

Got it. How important is the rollout of 5G base stations now? Because I've seen NXP increasingly describe the company more in terms of, this is our auto-industrial mix, right? So comms doesn't come in that center. I assume it's by design rather than.

Well, as you know, we had -- if you follow this long enough as you have, there are really 2 big players in the RF power marketplace, it was NXP legacy, let's call it, and Freescale legacy. We sold the legacy NXP business in 2016. We kept the Freescale business. It's a great business, but it's a very -- and this is not meant to be negative. It's a one-dimensional business. We sell a very unique product that goes into the towers on base stations, the high-performance RF power amplifiers. But there's no real other synergy with the rest of the business. And it's a great business. And it's a needed business. So there's a small number of sellers and a small number of buyers. So it's kind of a small tight marketplace. But it's extremely lumpy business and it's driven by carrier build-outs. And you can never truly forecast how those build-outs are going to happen.

Understood. And then on the mobile side, you mentioned, right, as the market obviously expects that there has been a slowdown on the Android, right? But isn't that also the place where you were expecting very high NFC attach rates?

Well, we said that the attach rate exiting 2021 was about 50%. In our internal models, we're not expecting a lot of fundamental handset unit growth over the next several years, if anything flattish to a little bit down. But what you do see in this is primarily a statement in the Android market is there are many different SKUs the Android OEMs have, both vertically meaning feature sets and horizontally along geographies. And they tend to add and subtract features depending on those 2 axises. And so we still think there's quite a bit of opportunity to win additional attach rate for mobile wallet. Think of that as kind of the core business of our mobile assets. We also have the ultra-wideband business, which is just starting to ramp. Now that is very linked to the security capability of the mobile wallet. And the use cases for ultra-wideband the kind of the killer application is secure car access. So using your phone as your virtual key get in and out of your car, manage your key, your car and things like that. But we're also seeing some very interesting ideas in the industrial IoT space for both commercial building access as well as home access and other, let's call them, more consumer-type in-home type of applications. We're very early days on ultra-wideband. But the ultra-wideband success is also predicated on the mobile wallet. I think they're kind of...

Are they kind of attached?

Yes, they're attached together. Well, ultra-wideband last year was about $80 million of revenue. We think it will grow to about $400 million by '24. It will primarily be driven by handsets in the short term and automotive as a secondary back to that 2- to 3-year design cycles in automotive. Any auto OEM who is doing a secure ultra-wideband-based secure car access system is working with NXP.

Got it. Terrific. Okay. Thank you so much, Jeff. Appreciate it.

Thanks. Appreciate it.

Thanks, everyone.

Thanks, everyone.